152 



CLIMATIC CYCLES AND TREE GROWTH 



Frequently an observer finds weak cycles which another observer, or even 

 the same observer at a different time, does not find. Such cycles (called 

 anomalous and not included in our lists of weak cycles) constitute a serious 

 matter for investigation. The amount of these cycles present are given 

 in table 2. 



It is significant to note that the out-in tests contain a much smaller per- 

 centage of anomalous cycles than the remaining kinds of tests; there is evi- 

 dently a dependence of one set on the other, as suggested already. The 

 similarity in the size of the remaining percentages indicates the absence of any 

 systematic factor; but an average of one anomalous cycle in every four seems 

 at first glance serious. It was found, however, that these are almost all weak 

 cycles of little weight, and random length, which largely cancel out when the 

 analyses in which they are found are summed into a periodogram. 



Another question which may be considered here is the difference in the 

 lengths of cycles analyzed at different times, as a possible function of the 

 length of the cycle. The available data are summarized in table 3. When 



Table 3 — Cycle differences with cycle length 



plotted there is evident a straight line relationship for each observer which 

 may be expressed as 



An = kn 

 where An = difference in cycle lengths, 

 k = constant for observer, 

 n = cycle length. 



From the above table we find the differences in cycle values, in per cent, 

 to be as follows: 



Pet. Av. Diff. 



The value of k is found to be 0.024 for AED and 0.020 for ES. 

 In summarizing this study of differences, we find that in spite of oppor- 

 tunities for disagreements the results of different observers are closely alike 



